Presentation of a data-driven, personalized nutrition risk assessment algorithm that utilizes food frequency questionnaires to identify dietary diversity clusters and examine their associations with health outcomes among middle-aged and older individuals in Taiwan. The study identified four dietary diversity clusters and found that the "most diverse" cluster had lower risks of developing hypertension, diabetes, and mortality compared to the "least diverse".